Artificial Idea | AI careers · practical prompts · no hype Monday, October 27, 2025 · Issue #25 · Jobs

The number that changes everything

Everyone read the headline. Almost nobody read page 47. Here is what was on page 47.

In the spring of 2025, PwC published its annual Global Workforce Hopes and Fears Survey, covering 56,000 workers across 50 countries. The report received substantial coverage. The coverage focused, as it almost always does, on the displacement numbers: the percentage of workers who fear job loss, the sectors facing the most acute automation pressure, the timeline projections for structural workforce change.

Page 47 of the full report received almost no coverage at all.

Page 47 contains a labour market analysis that PwC's own economists described as the most significant finding in the survey, one that they believed was being systematically underreported in the mainstream coverage of AI and employment. It shows, with granular sector and geography breakdowns, that job postings requiring AI-related skills grew by 7.5% in the twelve months covered by the survey, during the same period that total job postings across the same sectors fell by 11.3%.

That is not a marginal divergence. It is a 18.8 percentage point gap between the trajectory of AI-adjacent roles and the trajectory of the overall market. In a contracting hiring environment, one category of roles is growing at a significant rate. The professionals positioned to access those roles are operating in a fundamentally different labour market than those who are not.

The reason this finding received less attention than the displacement numbers is straightforward. Displacement is a more emotionally compelling story than bifurcation. Fear travels faster than nuance. But the professionals who make career decisions based on the fear story rather than the bifurcation story are optimising for the wrong variable, and the cost of that misoptimisation compounds over time in ways that become increasingly difficult to reverse.

What the 7.5% actually represents

The PwC figure requires careful reading because the category it describes, roles requiring AI-related skills, is broader and more heterogeneous than it initially appears.

The category is not limited to AI engineering roles, machine learning positions, or data science functions. PwC's methodology defines AI-adjacent roles as any position where the job description explicitly requires the ability to work with, direct, evaluate, or build on AI tools as a component of the primary job function. Under this definition, a marketing manager whose role requires AI-assisted campaign optimisation is an AI-adjacent role. A financial analyst whose role requires AI-assisted modelling is an AI-adjacent role. A legal professional whose role requires the ability to supervise and quality-check AI-assisted document review is an AI-adjacent role.

The 7.5% growth figure therefore does not represent a small and specialised corner of the labour market. It represents the leading edge of a broad structural shift in what employers are specifying when they hire across almost every professional function. The roles growing are not a new category of work. They are existing categories of work with an additional requirement attached, and the professionals who meet that requirement are accessing a labour market that is expanding while the one around them contracts.

A separate analysis by the Burning Glass Institute, published in June 2025, corroborates the PwC finding with additional granularity. It found that the salary premium for roles requiring AI skills relative to equivalent roles not requiring them had widened from 11% in 2023 to 19% in 2025, with the widening concentrated in professional services, financial services, and technology sectors. The premium is not static. It is growing, because the supply of professionals who genuinely meet the AI skill requirement is not keeping pace with the demand, and the market is pricing that scarcity accordingly.

The geography of the bifurcation

The PwC data includes geographic breakdowns that are relevant to this newsletter's readership and that the aggregate global figure obscures.

In India, the bifurcation is more pronounced than the global average. Total job postings in the sectors covered by the survey fell by 14.2% in India over the survey period, against the global average decline of 11.3%. AI-adjacent role postings in India grew by 9.1%, against the global average growth of 7.5%. The gap between contracting and growing segments of the Indian labour market is wider than the global norm, and it is widening faster.

The concentration of AI-adjacent role growth in India is also geographically and sectorally specific in ways that matter for individual career planning. The growth is concentrated in technology services, financial services, and professional services, specifically in roles that sit at the intersection of domain expertise and AI tool fluency. Roles requiring deep domain knowledge in a professional function combined with demonstrated AI capability are growing faster than either domain-only or AI-only roles, which suggests that the most valuable position in the Indian professional market is not the AI specialist but the domain expert who has developed genuine AI fluency.

This finding aligns with the argument made in Issue #5, that the combination of deep domain knowledge and AI fluency is the combination the market is paying a premium for, and it provides empirical support for that argument in the specific Indian context where this newsletter's readership is concentrated.

What employers are actually specifying

The Burning Glass analysis includes a breakdown of the specific AI skills most frequently appearing in job descriptions across the roles contributing to the 7.5% growth figure. The breakdown is instructive because it reveals that the skills in demand are not primarily technical.

The most frequently specified AI skill, appearing in 43% of AI-adjacent job postings, is the ability to evaluate and quality-check AI-generated outputs. This is not a technical skill. It is an analytical one, requiring the professional to assess whether an AI output is accurate, appropriate, and fit for the purpose it will be used for. The critical thinking capability discussed in Issue #9 is not an abstract professional virtue in the context of this data. It is the specific skill most frequently cited by employers hiring into the fastest-growing segment of the labour market.

The second most frequently specified skill, appearing in 38% of postings, is prompt engineering for professional applications, defined as the ability to direct AI tools to produce outputs relevant to specific professional contexts. This is the skill this newsletter's Thursday issues have been building since Issue #2.

The third is AI-assisted research and analysis, appearing in 31% of postings. The fourth is AI tool integration into existing workflows, at 27%. The fifth is the ability to communicate AI-assisted findings and recommendations to non-technical stakeholders, at 24%.

Every one of these skills is domain-agnostic. They transfer across industries and functions. They are learnable without a technical background. And they are, collectively, the specific capability set that the fastest-growing segment of the professional labour market is currently paying a premium to access.

The compounding dynamic

The salary premium data from the Burning Glass analysis reveals something that is easy to miss in a single-year snapshot but becomes significant when viewed as a trajectory. The premium grew from 11% to 19% in two years. If that trajectory continues at a similar rate, the compensation differential between AI-fluent and non-AI-fluent professionals in equivalent roles will exceed 30% within three years.

Compounding salary differentials produce outcomes that are difficult to reverse. A professional who is earning 19% more than their non-AI-fluent peer today is accumulating savings, building a financial cushion, and developing a track record at a higher compensation level that anchors their next negotiation. The peer earning less is not just behind today. They are in a position that becomes harder to close each year the differential persists.

This dynamic is not unique to AI. It is the standard mechanism by which skill premiums operate in professional labour markets. What is specific to the current moment is the speed at which the differential is widening and the breadth of the professional population it is affecting. Most skill premiums are narrow, affecting specialists in specific technical domains. This one is broad, affecting knowledge workers across almost every professional function, and it is widening at a rate that makes the decision to engage with it or not increasingly consequential with each passing quarter.

The action

Pull up your current resume or LinkedIn profile. Search it for the five skills most frequently specified in AI-adjacent job postings according to the Burning Glass analysis: output evaluation, prompt engineering, AI-assisted research and analysis, workflow integration, and communicating AI-assisted findings to non-technical stakeholders.

Count how many of them are represented, with specific evidence, in your current professional profile. The number you arrive at is a rough proxy for your current position on the bifurcation curve. It is also a prioritised list of what to work on next, ordered by market demand rather than personal preference.

Market demand is not the only variable worth optimising for in a career. But it is a variable that compounds in ways that personal preference alone does not, and understanding where you currently sit relative to it is a precondition for making an informed decision about what to do next.

Thursday we are giving you the prompt framework that turns this kind of skills gap analysis into a structured, prioritised development plan with specific milestones, realistic timelines, and a mechanism for tracking progress that does not rely on willpower alone.

The plan is only as useful as the honesty of the inputs. Thursday's framework is designed to make the honesty easier.

— The Artificial Idea team

You're receiving this because you subscribed to Artificial Idea · Unsubscribe